Photoionization of diarylmethyl radicals in acetonitrile and alcohol

Photoionization of diarylmethyl radicals in acetonitrile and alcohol-water: laser flash production of diarylcarbenium ions. Joaquim L. Faria, and S. S...
0 downloads 0 Views 867KB Size
1924

J. Phys. Chem. 1993,97, 1924-1930

Photoionization of Diarylmethyl Radicals in Acetonitrile and Alcohol-Water: Laser Flash Production of Diarylcarbenium Ions Joaquim L. Farial and S. Steenken' Max-Planck Institut f i r Strahlenchemie, 0-4330 Miilheim, Germany Received: November 1 1, 1992

Diarylmethyl radicals Ar2CI-P were produced by 248-nm laser photolysis (20 ns) of diphenylmethane, diphenylmethanol, and diarylmethyl halides in acetonitrile and water:alcohol mixtures. A few microseconds after their generation, the radicals were subjected to a laser pulse (20 ns) of 308-nm light. As products of the photolysis of the radicals, the electronically excited radicals, Ar2CH** (DI state), and the corresponding cations, Ar2CH+, were identified. The electronically excited radicals decay with emission of light ( 4 5 0 nm in the case of Ph2CH'*) to their ground state (Do),whereas the cations react with solvent or (other) nucleophiles. The formation of cation from Ar2CH' requires two photons (is biphotonic) and proceeds by photoionization of Ar2CH'*. The photoejected electrons were directly observed in aqueous alcohol. In acetonitrile, the yield of photoionization of Ar2CH' (as measured 1 2 0 ns after the 308-nm pulse) can be increased by the electron scavenger n-butyl chloride, which does not react with Ar2CH'*. In contrast, C C 4 , which does react with Ar2CH'*, does not lead to an increase in cation yield.

Introduction

In many photochemical reactions, free radicals are important intermediates, resulting from (homolytic) abstraction, fragmentation, and ionization processes. Under conditionsof low intensity of exciting light, in solution the stationary concentration of the radicals is typically M, due to their usually large rates of chemical reaction (radical-radical termination rates IO9-lO1O M-I 5-I). As a result, absorption of light by the radicals can be neglected compared to that by the parent compounds, even if the extinction coefficients of the radicals are up to 103-104 times higher. However, the situation is different if high intensity (pulsed) methodssuchas,e.g., laser photolysisareusedtogenerate radicals which permit initial radical concentrations in the pM to mM range to be easily produced. Under these conditions, the photolysis of radicals cannot be neglected and it becomes necessary to understand their photochemistry in order to obtain a complete picture. The photochemistry and photophysics of diary124 and triarylmethyl' radicals in solution have previously been investigated using the Ytandempulse" techniques and other9-" methods. In the more recent studies, the diphenylmethyl (DPM) radical was produced by irradiation of diphenylmethyl chloride with an electron pulse5 or by photolysis of 1,3-di~henylacetone,~ and photolyzed with 347- or 337-nm light. The excited state of the DPM radical thus produced is quite long-lived (lifetime 280 ns) and highly fluorescent (@f = 0.30) from its lowest excited state (DI).5 The fluorescence quantum yield was found to be virtually independent of the nature of the solvent. However, in several polar solvents, after decay of the excited state, the recovery of the ground-state DPM radical is not complete, which means that an irreversible photochemical reaction occurs. In line with this is the fact that the fluorescencequantum yield decreases at higher laser light intensities, from which it was concluded that the photochemical process leading to incomplete ground-staterecovery arises from the absorption of a second photon by the first excited doublet state, in a consecutive process. It was tentatively suggested that upon further excitation of the DI state homolytic cleavage of the C-H bond occurs with formation of the corresponding carbene.sb However, since the carbene would abstract a hydrogen atom from the solvent giving the ground-state radica1,ba this mechanism does not account for the incomplete recovery of the ground-state radical. 0022-3654/93/2097-1924$04.00/0

We have previously found that it is possible to ionize a neutral radical, such as triphenylmethyl radical, in a polar solvent, such as water, and to make the resulting carbocation visible.' In the present work we demonstrate that the same reaction takes place also with (substituted) diphenylmethyl radicals,l2 Le., that the permanent destruction of the ground-stateradical under conditions of high light intensities is due, at least in polar solvents, to its photoionization. On the basis of the ionization potentials of diarylmethyl radicals,13this type of reaction is thermodynamically possible with 308-nm light (quantum energy 4.0 eV), if the solvation energies of the ions are >3 eV.14 Experimental Section Para-substituted diphenylmethyl (benzhydryl) chlorides were prepared from the corresponding alcohols (from Lancaster) by treatment with HC1 in CH2C12,Is and they were purified by fractional distillation or recrystallization. Benzhydryl bromide (from Aldrich) was zone refined toa purity >99%. Benzophenone (>99%) and diphenylacetic acid (>98%) were from Fluka, diphenylmethane (>99%) from Aldrich. Acetonitrile (=AN), dichloromethane,isopropanol, and tetrachloromethane (Merck) and n-butyl chloride (Fluka) were spectroscopic grade. Solutionsof the substrates (A/cm 0.2-2) were flowed (rates = 1-6 mL/min) through a 2 mm (in the direction of the laser beam) by 4 mm (in the direction of the analyzing light, 90° geometry) Suprasil quartz cell and photolyzed with 20-ns pulses (denoted as P1,l-100 mJ, as measured at the position of the cell, typical values are 10 mJ) of 248-nm light from a Lambda Physik EMG103MSC excimer laser, and the time-dependent optical changes were recorded with Tektronix 7612 and 7912 transient recorders interfaced with a DEC LSI 11/73+ computer, which alsocontrolled theother functionsof theinstrument and performed on-line analysis of the experimental data. A few microseconds after the first pulse, the solution was subjected to a second 20 ns laser pulse (denoted as P2, 1-60 mJ, at an angle of -loo to the analyzing light) of 308 nm from a Lambda Physik EMGl50E laser. To convert the mJ numbers given in the text or in the Figures into mJ/cm* at the position of the cell, the numbers given for the 248-nm pulses should be multiplied by 2, those for the 308 nm pulses by 0.9. The delay between the first pulse and the second could be varied from 10ns to the ms range by a computercontrolled delay generator. Time-dependent absorption spectra AA vs X or kinetics refer to the situation after the first (Pl, the

-

0 1993 American Chemical Society

The Journal of Physical Chemistry, Vol. 97, No. 9, 1993 1925

Laser Flash Production of Diarylcarbenium Ions

t 4

a

0 200

400

300

600

500

700

Ilnm

Figure 1. Time-resolved absorption spectra generated by 248-nm photolysis (14 mJ/pulse) of 1.1 mM Ph2CHBr in acetonitrile, saturated with argon, recorded 50 ns (full circles) and 2.5 ps (open circles) after the pulse. Inset: Spectra from the same solution upon two-pulse laser photolysis (PI = 248 nm, 14 mJ, P2 = 308 nm,60 mJ, delay = 3 ps), recorded at 85 ns (full squares), 270 ns (open squares), 470 ns (full circles), and 2.5 1 s (open circles) after P2.

photolyzed with 248-nm light (pulse 1, P1) in acetonitrile, water, ethanol,or mixtures ofthese solvents. The diphenylmethyl radical, PhZCH' (DPM'), was formed, as judged by its characteristic absorption spectrum with A, at 330 nm. The formation of DPM' requires two quanta, as concluded from the quadratic dependence of the DPM' yield (measured as L4 at 330 nm) on the laser light intensity I ( L 4 = const.12). With PhZCHCO2-in aqueous solution,1g the formation of e-aq was observed on photolysis, showing that DPM' formation is initiated by photoionization. In agreement with the idea of DPM' production proceeding via the radical cation of the aromatic (eqs 6,7) is the fact18that the radicalcan be generatedindependently by reaction of the one-electron oxidant SO4'- with PhzCHC02-. From the observation that there is no delay in the formation of DPM' after the 20 ns pulse, the rate of deprotonation of Ph2CHz'+ or of decarboxylation of PhzCHCOZ-*+ is >5 X lo7 s-I. Similar conclusions have previously been drawn for phenyl acetates20J and triphenyl acetate:' PhzCH, Ph,CHCO;

"synthesis"6) or after the second (P2, the "photolysis"6) pulse, as indicated in the figures or text.

Results and Discussion A. Formationof DiruylmethylRadicals. Diarylmethyl radicals were produced basically by two different processes:I6 Photohomolysis of ArzCH-Y (Ar = para-substituted benzene, Y = Hal, OH) by 248 nm light: Ar,CH-Y

-

Ar,CH*

+ Y'

(1)

or side-chain fragmentationi7of diarylmethyl radical cations Ar2CH*+Y (Y = H, COZ-), formed by biphotonic (A = 248 nm) ionization from diarylmethane derivatives: Ar,CRY

+ 2hul-e-

-

Ar,CR'+Y Ar,CR'

-+

Y+, R = H, OH (2)

The advantage of using these simple, nonconjugated "alkylbenzenes", which have a strong *** absorption at 248 nm, the wavelength of the "synthesis pulse" (Pl), is that they essentially have no absorption at 308 nm, the wavelength of the "photolysis pulse" (P2). Photolysis at this wavelength, therefore, leads to excitationof only the products of the 248 nm photolysis (provided, of course, that they absorb at 308 nm), but not of their parent compounds. Pbotohomolysis (Eq 1). Diarylmethyl Halides. On 248-nm photolysis in acetonitrile (AN), diarylmethyl halides undergo efficient homolysis (eq 3, @I,,, =,0.1-0.39) and heterolysis (eq 4, ahet= 0.05-0.32) to yield the corresponding radicals and cations.I8 The latter react rapidly with acetonitrile (lifetimes 5 ps). Since the electronically excited radical~s,~ that result from biphotonic excitation at 248 nm in these systems1*also have a after a few microseconds, the ground short lifetime ( state radicals are the only transient species remaining (see Figure 1). In the absence of oxygen, their lifetime typically is 1 ms. In Table I are listed some spectroscopic and kinetic data of the transients from the benzhydryl halides used. ArzCH* + X*

-

Ph,CH'

+ H+ + e-solv

+ 2hu/-e- or +SO,'-/SO,2-

-

-

Ph,CH' Ph,CH'

+ R+;

+ CO,

(5)

(6)

R = H, C O G

k 15

X

lo7 s-' (7)

Diphenylmethanol. The spectrum obtained upon photolysis of diphenylmethanol in acetonitrile (Figure 2a) is very similar but not identical to that of the DPM radical. In particular, there is a weak band at 550 nm which is not from Ph2CH'. This band indicates the presence of the a-hydroxydiphenylmethyl radical which can be obtained quantitativelyby, e.g., hydrogen abstraction by excited benzophenone from a hydrogen donor such as 2-propanol. The spectrum (Figure 2b) observed under these conditions (lifetime of the excited triplet stateZZof benzophenone 46 ns) matches that measured in acetonitrile (A, = 328 nm, t = 3.0 X lo4 M-I ~ m - l ) .There ~ ~ was no evidence for formation of the diphenylmethyl cation in acetonitrile, which means that photoheterolysis does not occur in this solvent. The absorption band of a-hydroxydiphenylmethyl radical at 328 nm (Figure 2b) is very similar to that of the diphenylmethyl radical at 330 nm (Figure 1, 2.5 ps after the pulse). It was found that the signal intensity as measured at 330 nm increased with the laser intensity I in a supraproportional way, the relation being hA = constel + consV-12. This suggests that the species absorbing at 330 nm is (are) produced in a monophotonic and in a biphotonic way. These processes are assigned to photohomolysis of the C-OH bond (eq 8, one photon required) and photoionization (eq 9, two photons necessary) followed by deprotonation of the resulting radical cation (eq 10): Ph,CHOH

E Ph,CH' + OH'

Zhu

Ph,CHOH Ph,CHOH'+

-

Ph,CHOH'+

Ph,C*OH

+ H+,

(8)

+ e-

k 15

X

(9) lo7 s-'

(10)

(3)

Ar2CHX Ar2CH* + X-

Ph2CHR'+

+ 2hu

(4)

Diphenylmethanol. For the results, see eq 8. Side-ChainFragmentationof Radical Cations (42 ) . Diphenylmethane and Diphenylacetic Acid. These compounds were

Photohomolysis (eq 8) could be very clearly seen in the unpolar solvent cyclohexane,where ionization is certainlyvery unfavorable. The spectrum observed in this solvent at 4 ps after the pulse was identical with that in Figure 1 (open circles). The occurrence of ionization (reaction 9) was tested by photolyzing PhzCHOH in aqueous solution which led to the characteristic absorption of the hydrated electron, e-aq,in a biphotonic reaction. Reaction 10 is

1926 The Journal of Physical Chemistry, Vol. 97, No. 9, 1993

Faria and Steenken

TABLE I: Absorption Maxima of Diuylwtbyl Radicrrle and Cations (4-RCfi)&H. the G r d and Excited State Radical8 and for Decry of tbe Cation8 in Acetonitrile

H, H CH,, CH3 CHjO, CH3O CI, c1 CI, H

330 335 350 340 334

1 .I (0.6)b 1.9 2.0 1.Y 1.2

355,435c 365, 474c 380" 310, 474c 361,44Y

Rate Constants for Reaction with 02 of

3.6 4.8

8.1b 12

5.8

11

6.5 5.3

9.u

9.6

435 465 500 415 451

2.4 0.22 e

3.0 3.3

Similar to the numbers given in ref 18 and references therein. In cyclohexane, from ref 6a. Broad band of weak intensity ( - 3 ~ 4 0 % of that at the lower A). d Band at longer X not measurable. Cation decays bimolecularly by combination with anion, see ref 18. /Taken from ref 18. 0

(formation) peaking at 350 nm; c) a strong positive band with at 435 nm; and d) a weak negative signal in the 550 to 600 nm region. With the exception of the negative absorption at 330 nm, the optical density changes return to zero within 3 ps. Identiflcation of the P2-Induced Phenomena. A "negative absorption" after P2 means that more light is seen by the photomultiplier relative to the situation before P2. This can be caused by a larger transmittance of the solution, Le., by Figure 2. (a) Absorption spectrum measured 4 p s after 248-nm photolysis photobleaching, or it can be due to light emission (fluorescence). (PI, 24 mJ) of 3.6 mM Ph2CHOH in acetonitrile, saturated with argon. Since the negative absorptions in the 330-350-nm region (for (b) Absorption spectrum recorded 425 ns after 248-nm photolysis (PI, Ph2CH' and the substituted benzhydryl radicals) are perfect 8 mJ) of 32 pM PhZCO in 2-propanol, saturated with argon. mirror images of the absorptionspectra of the radicals (see Figure 3, inset a, for the case of (4-ClC6H4)2CH'), these negative in line with the dramatically enhanced Br6nsted acidity of radical absorption changes are identified in terms of depletion of the cations2426 compared to their parent compounds. ground-state radicals. To summarize, in the polar solvent acetonitrile diphenylmethanol undergoes a monophotonic C-OH homolysis and a biphotonic In contrast, the negative signals observed at 5 0 0 6 0 0 nm can ionization, whereas in a nonpolar solvent such as cyclohexane not be explained in terms of bleaching since at these wavelengths only homolysis is observed. the radicals have no absorption (see Figure 1). Therefore, they Diphenylmethanol was also photolyzed in the non-nu~leophilic~~ must be due to emission induced by P2. This was demonstrated and acidic solvent 1,1,1,3,3,3-hexafluoroisopropanol(HFIP, pK, by working without analyzing light, which made the detection of = 9.39, which has been shown to be able to protonate substituted the emission much easier. On the basis of the similarity of the ben~enes,2~J'J styrenes," and phenyla~ctylenes~~ upon their spectra of these emissionswith those6bpreviously measured, they electronicexcitation. The spectrum observed after the pulse was are identified as fluorescence from the first electronic excited that of the benzhydryl carion, Ph2CH+. It was formed monostate (Dl) ofthebenzhydrylradicals. In support ofthisassignment photonically and there was no evidence for production of the is the fact that the fluorescence decay rates are exactly the same corresponding radical, as supported by the finding that the as those measured for decrease of the excited state absorptions spectrum was the same in the absence or presence of oxygen. This or for recovery of the ground-state radicals (vide infra). means that in the only weakly polar solvent HFIP (e = 16.6)33 The most conspicuousP2-induced absorption changesare those photoheterolysis is the predominant reaction, in contrast to the due to the electronically excited diarylmethyl radicals in the 330 situation in the much more polar solvent acetonitrile (e = 37.5). to 370 region and those of the cations between 435 and 500 nm. The increase in polarity in going from HFIP to acetonitrile is As previously ~ h o w n ,the ~ ~cations , ~ ~ have a very high reactivity thus not sufficient to compensate for the lack of acidity of with halides (k = (1-2) X 1OIo M-l s-l). Therefore it is possible acetonitrile. Thus, acidity seems to be the decisive property in to quantitatively scavenge the cations with relatively small enabling heterolysis to occur. The easiest way to understand this concentrations of halides. Alternatively, alcohols or water can is in terms of protonation of the OH group, caused by its increased be used as reactive nucleophiles.~*J7In experiments of this type, basicity in the excited state.34 As a result, the bad leaving group it was found that after complete removal of the cations, there OH- (pKa(H20) = 15.7) is converted into the excellent one H2O remained absorptions in the 435-5Wnm region. An example (pKa(H30+)= -1.7): for this is shown in Figure 3, inset b. These absorptions decayed -(CF>)2CH& with the same rates as those in the 350 to 370 nm range, as those Ph2CH-OH* (CF,),CHOH of the ground-state recovery, and as those of the fluorescence in the 500 to 600 region (see Figure 3, insets d-g). The absorptions Ph,CH-O+H2 Ph2CH+ H 2 0 (1 1) could be removed with oxygen, again with the same rates as those Using the number 4.4 X 104 M-1 cm-I for the c of Ph2CH+at in the lower wavelength region. The conclusion is thus that the 435 nm and aqueous K2S208 for actinometry35 with @c(S04*-) (broad) absorption bands that extend from 400 to 500 nm under = 2770 M-I cm-I at 445 nm,36the quantum yield for formation conditions of cation scavenging are due to the excited radicals. of the benzhydryl cation was determined to be 0.013. These absorptions have been seen previously.6a The excited state B. 308-nmPbotolysisof Muylmetbyl Radials. Diarylmethyl radicals, Ar2CH**, have a much higher reactivity with oxygen Halide Precursors. In Figure 1are shown the absorption spectra (for the case of (4-ClCsH4)2CH'*, k = 9.4 X 109 M-I s-l) than of Ph2CH' (A, 330 nm) and of Ph2CH+ (A, = 435 nm) the ground-state (k = 1.3 X lo9 M-I S-'),,~ and they can be resulting from 248-nm photolysis (pulse 1, P1) of Ph2CHBr in scavenged with 1,3-cyclohexadiene,18 which shows no reactivity acetonitrile ( q s 3 and 4). It is evident that after 2.5 ps, the with the ground-state radicals. cation has disappeared (by reaction with solvent),I8whereas the Finally, the pronounced bell-shaped bands with the maxima radical concentration has hardly changed. At 3 ps after P1, a pulse of 308 nm light was applied (pulse 2, P2). As seen in the at 435 to 500 nm are identified as the ground-state (substituted) inset, P2 leads to: a) a "negative absorption" (depletion) with a benzhydryl cations. Not only are theabsorptionspectra identical maximum at 330 nm; b) a positive optical density change with those previously deter~nined,~ for the cations in acidic media, Amax

+

- +

Laser Flash Production of Diarylcarbenium Ions

A 4: Q

0200

3W

400

500

+

600

m2m

700 Mo

The Journal of Physical Chemistry, Vol. 97, No. 9, 1993 1927

300

1

tnm

U

a 0

200

300

500

400

600

700

A/nm Figure 3. Time-resolved absorption spectra generated upon two-pulse laser photolysis (P1 = 248 nm, 10 mJ, P2 = 308 nm, 60 mJ, delay = 3 ps).of 0.2 mM (4-CIPh)zCHCl in acetonitrile, saturated with argon, recorded 75 ns (full circles) and 3.0 p s (open circles) after P2. Insets: (a) Absorption spectrum recorded at 2.5 ps after P1 (full circles), and spectrum measured at 5 p s after P2 (delay 3 ps) (open circles). (b) Spectrum measured at 130 ns after P2 (delay = 3 ps) in (4-ClPh)2CHCI in acetonitrile containing 4 mM (n-Bu)rNCl, saturated with argon. (c) Spectrum recorded at 220 ns after P2 (delay = 3 1 s ) in (4-CIPh)2CHCI in acetonitrile, saturated with an argon:oxygen mixture (1.7 mM 02). (d)-(g) Kinetic traces recorded in the presence of 4 mM (n-BuhNCl showing the recovery of the ground state radical (d), the decay of the excited state radical (e and f), and the decay of emission from the excited state radical, all with k = 6.4 X lo6 s-l.

or upon photoheterolytic formation in acetonitrile,18but also the electrophilicreactivities are the same as those18J7measured after heterolyticproduction. Concerningtheir absorption spectra,there is overlap with the spectra of the electronically excited radicals, as mentioned above. Thesecan, however, be selectively scavenged by small concentrations of molecular oxygen (12 mM; with 0 2 saturated solutions ([OZ] = 8 mM), the ground-state radical precursor to the cation will be trapped), leading to a pure spectrum of thecation (see Figure 3,inset c, for thecaseof (4-CIPh)2CH+). This procedure utilizes the difference of reactivity between the excited and the ground-state radicals with molecular oxygen (for therateconstants,seeTableI). Notethat ininsetc theamplitude of ground-state radical bleaching is similar to that of cation formation. Under the condition that the reaction of the excited state radical with oxygen leads to recovery of the ground-state radical (by energy transfer) and that the extinction coefficients of the two species are comparable,5.68 this result means that the depletion of radical is caused by its conversion into cation (ionization). Diphenylmethyl halides substituted at the para-position of the ring(s) by only one C1, and by Me and M e 0 were also studied (for identification of the systems, see Table I). In all cases, after P2, very intense negativeabsorptions (which are due to P2-induced depletion of the corresponding radicals, for the A, and c values, see ref 18) were seen between 330 and 350 nm, peaks with absorption maxima between 340 and 360 nm (due to the electronicallyexcited radicals), intense absorptions between 435 and 500 nm (representing the cations),I8 and very weak but observable negative changes above 500 nm (resulting from fluorescenceof the excited radicals). For an example, scc Figure 3. Diphenylmethane and Diphenylacetic Acid as Precursors. As pointed out above, the benzhydryl cations are derived from the radicals by photoionization. In this process, by definition, an electron is formed, and it is therefore tempting to try to demonstrate directly the production of the electron. To see this species, the presence of electron scavengers should be avoided, and it is for this reason that diphenylmethane, diphenyl acetate,

IA

I 200

.

,

300

.

a

.

400

* 500

Unm

.

.

600

a.

Y

I I . l . I 700 200 300 400

I Mo

600

700

Unm

Figure 4. (a) Spectrum recorded at 150 ns after P2 (40 mJ, delay 3 ps) in a 3 mM solution of Ph2CH2 in AN saturated with Ar. (b) Spectrum recorded at 70 ns after P2 (60 mJ, delay 2 p s ) in 3 mM Ph2CH2 in W:EtOH (1:3 v/v). Full circles: Ar saturated. Open circles: N20 saturated.

and diphenylmethanol(seelater), which all have low rateconstants for reaction with the electron, are suitable precursors for the benzhydryl radical from which the electron is to be photoejected. As seen in Figure 4a, the absorption spectrum observed upon double pulse laser photolysis of Ph2CH2 in acetonitrile solution saturated with Ar, recorded after P2, is qualitatively the same as that (see Figure 1) observed with Ph2CHBras precursor. Even the fluorescence emission of PhzCH" is visible at -550 nm. The fact that the amplitude of the cation (at 435 nm) relative to that of the electronically excited radical (at 350 nm) is smaller than in Figure 1,inset, is due to the intensity of the 308 nm pulse being less than in the former case (see photonity data in section B). It is known that in acetonitrile the electron, which exists as the dimer radical anion (AN)2*-,40has only a very weak absorption between 400 and 700 nm,41 and is therefore difficult to detect in this solvent. To see the electron, it is necessary to use solvents in which e-=l, absorbs strongly. This is the case with (aqueous) alcohols. In water:ethanol 1:3, the spectrum, obtained after P2 under the same conditions as in the case of acetonitrile solutions, is characterized by the bleaching of PhZCH', the absorption and emission of its electronically excited state, and a new band with a maximum above 550 nm, but no cation band (Figure 4b). The

Faria and Steenken

1928 The Journal of Physical Chemistry, Vol. 97, No.9, 1993

12

0 d 0.48 - 1

0.20 m

I (mJ/ pulse)

200

I

I

I

I

300

400

500

600

700

300

600

400

d

I (mJ/ pulse)

Figure 6. (a) Dependence of yield of excited state radical (open squares) and of cation (full squares) on the intensity (mJ/pulse) of P2 (delay 3 ps) in 0.3 mM (4-CIPh)$HCI in AN saturated with Ar. (b) Open squares: Yield of (measured at 680 nm 30 ns after P2) as a function ofI(P2), theintensityofP2,ina 1:3 (v/v) water:EtOHsolutioncontaining 1.1 m M Ph2CH2. Full squares: Yield of e-solvas a function if I(P2)2.

A/nm Figure 5. Spectrum observed after P2 (60 mJ, delay 6 ps; P1 24 mJ) in argon saturated 3.6 mM PhzCHOH in AN at 125 ns (full circles) and 250 ns (open circles).

550 nm band can be efficiently removed by N20, 02, and 2-chloroethanol or other halogenated hydrocarbons, with bimolecular rate constants similar to thosede~cribed~~ for e-aqreactions. This band is therefore identified as due to e-solv. In agreement with this is the fact that the subtraction of the spectra obtained in absence and in presence of N 2 0 results in a spectrum comparable to the absorption spectrum of the hydrated electron. The fact that the cation band cannot be observed at 220 ns after P2 is due to the high reactivity of the nucleophilic solvent. The cation lifetime in water:EtOH 1:3 can be calculated to be a80 ps, using the bimolecular rate constants 1.2 X lo8 M-I s-I for reaction with water and 8.5 X 108 M-I s-l for reaction with EtOH.18 The results obtained with Ph2CHC02- in purely aqueous solution are similar to those described above for Ph2CH2 in W:EtOH. Benzhydrol Photolysis. As reported in section A, 248-nm photolysis of diphenylmethanol(benzhydrol) in acetonitrile leads to the benzhydryl and to the a-hydroxydiphenylmethyl radical (eqs 8-10). In Figure 5 are shown the spectra observed after P2 in a diphenylmethanol solution in acetonitrile. The electronically excitedbenzhydrylradical is clearly recognizableby its absorption at 350 nm and its emission at 550 nm, and also the benzhydryl cation is present (A, = 435 nm). The difference to the situation with the other ground-state radical precursors is the wavelength region between 240 and 350 nm. Here, the negative band is much broader, and at 252 nm, there is a pronounced positive band which is very long lived (1ms). On the basis of its shape and its lifetime of > 1 ms, this band is assigned to a stable product, i.e., to benzophenone, known to be formedM by photoinduced rupture of theO-H bond in the a-hydroxydiphenylmethylradical:

hv

Ph,C'O-H

Ph2C0

+ H'

(12) The formation of Ph2CH'* and Ph2CH+ by P2 is confirmation for C-OH homolysis by P1 (eq 8), the productionof benzophenone by P2 support for generation and deprotonation of radical cation ( q s 9 and 10). Photon Demandfor Formation of Ar$H'* and ArzCH+. With diarylmethyl halides as precursors, at constant intensity of P2, the photochemical yields of the transient products of P2, i.e., Ar2CH'* and Ar2CH+,were proportional to the intensity of P1, the synthesis pulse. This result is in agreement with expectation for the monophotonicls C-C1 bond homolysis induced by P1. Concerning P2, at low intensities the yield of Ar*CH'* was found to increase linearly with the intensity of P2, I(P2). However, at 1 1 0 mJ/pulse the yield starts to level off and at >30 mJ it

decreases again (for an example involving (4-ClC6H4)2CH', see Figure 6a). The initial linear portion in the dependence on I(P2) is as expected on the basis of the observed18quadratic dependence on I(P1) from 248 nm photolysis (Le., it takes two photons to produce ArzCH'* from Ar2CH41, but only one from ArCH'). Concerning the formationof cation from the ground-state radical, it is interesting that two photons are rquired to achieve this. This is visible with the initial portion of the yield vs. intensity plot, which is curved upward. After this, there is a quasi-linear portion, due to incipient saturation effects. It is interesting that the onset of saturation in the formation of Ar2CH'* occurs in the intensity range (-1C30 mJ/pulse) where the increase in the formation of Ar2CH+ is most visibly curved upward. Also relevant is that the yield of Ar2CH'* goes through a maximum, i.e., that at higher laser intensitiesbleaching of ArZCH**occurs. This indicates that cation formation is at the expense of the excited radical. The conclusion is then that the cation is formed by absorption of a 308 nm light quantum by the excited radical. This process is apparently quite efficient. The susceptibilityto photolysiswithin the same pulse is a consequence of the long lifetime (100-330 ns)18 of Ar2CH** (DI~ t a t e ) . If~ . ~ the lifetime was in the low ns range, as typical with excited singlet states of aromatics, the chances for photochemicaltransformation would be much les43 To check the photonity of 2 found (from the quadratic portion of the yield vs intensity plot) for formation of cation, the yield of electrons produced by P2 was measured as a function of its intensity, using Ph2CH2 in water:ethanol as precursor for Ar2CH'. The results, presented in Figure 6b, indicate that two photons are required for ionization of the radical, which is in support of the data for cation formation. The results presented in the previous paragraph show that ionizationofAr2CH' takes place by photoexcitationof theexcited radical, Ar2CH'*, Le., by a consecutive biphotonic process. In order to further investigate the nature of the ionization process and to judge the potential influence of the fate of the ejected electron on the yields of the transient products, the effects of electron scavengers was studied that react with the electron irreversibly, such as chloroalkanes, which undergo dissociative electron capture. In all cases, the effect of the scavengers on the lifetime of Ar2CH** and on its yield and on that of permanent depletion of the ground-state radical (measured at >3 ps after P2) was also determined. It was found that none of the scavengers decreased the lifetime of Ar2CH**, except CC4, for which the rate constant for reaction with Ar2CH'' was measured to be 3.8 X 108 M-I s-I. This number is similar to that reported6*for the same reaction in methanol. However, all scavengers increased the degree of permanent depletion of ArICH' (by the factor -2). Results obtained for the case of Ph2CH' with n-butyl chloride (n-BuC1) and C C 4arc contained in Figure 7a and b, respectively. In all cases the data have been corrected for influences of solvent on the yield of ground-state radical produced by P1.

Laser Flash Production of Diarylcarbenium Ions 0.2

,

,

0.0

0.2

The Journul of Physical Chemistry, Vol. 97, No. 9, 1993 1929

,

,

LO

SCHEME I: Energy Diagram for Ionization of Pb2W in CHjCN

-01

{ 01

E

Q

- 02

t

7.3

Ph2CHC

+

e-

s

-03

00

0

nC,H.UWroll

1

2

3 4 CCI,P* rol I

6

0

Figure 7. Effect of n-BuCI (a) and of CC14 (b) on the photochemical yield of Ph2CH'* (triangles), of Ph2CH+ (squares), and of permanent depletion of Ph2CH' (open circles), resulting from P2 (3040mJ) and measured as initial changes in A/cm at 350,435 (at -30 ns), and 330 nm (at 4-6 ws after P2), respectively, in a solution of 0.3mM Ph2CHBr in AN, saturated with argon.

As seen, the effect of n-butyl chloride (Figure 7a) on the photochemical yield of cation is different from the effect of CC14 (Figure 7b). With n-BuC1, the yield of cation increases with [n-BuCI] to a maximum after which there is a decrease to a lower value in pure n-BuCI as compared to AN. The yield of Ph2CH'* is rather independent of [n-BuCl]. whereas the permanent depletion of the ground-state radical increases with increasing [n-BuCl], without going through a maximum as in the case of the cation. Avery similar effectof n-BuCI on ground-state,excited state, and cation yields was observed with the (4-MeC6H4)2CH system (not shown). From Figure 7b it is evident that evenvery smallconcentrations of CCld drastically decrease the yields of excited radical,44while leading to a pronounced increase in the ground state permanent depletion. In contrast to the effect of n-BuC1, there is no increase in the yield of cation, on the contrary, the cation yield decreases in a way similar to that of Ph2CH**. The conclusion is thus that neither Ph2CH** (D1) nor the state produced from the D1 state by absorption of a 308 nm photon (denoted as Ph2CH'**) react with C C 4to give cation measurable at 1 2 0ns after P2. However, there must be a product-forming reaction between excited state radical and CC4, since there is a strongly increased yield of permanent depletion of ground-state radical with increasing [CCh]. This reaction cannot be electron transfer with subsequent solvation and separation of the ions formed, unless one assumes that CC14reduces the lifetime of Ph2CH+drastically. This is not at all the case, as was found by producing the cation by photoheterolysis (eq 4) with P1. No effect of C C 4on the cation lifetime was seen up to 5 vol I. It is therefore suggested that the reaction with Ph2CH**or Ph2CH***involves C1-abstraction. An alternative is electron transfer followed by quantitative collapse in the solvent cage of the hypothetical ion pair The effect of CH2C12was also studied (with the PhzCH and the (4-CIPh)2CH systems): CH2C12 had no effect on the lifetime of Ar2CH'*, it increased the yield of permanent depletion of Ar2CH*,but it did not lead to an increase of the yield of Ar2CH+. In pure CH2CI2 the yield of cation (at 120 ns after P2) was much less than in AN. Concerning the effect of n-butyl chloride (Figure 7a), after the initial increase in the cation yield, which can most easily be explained in terms of scavenging of the multiply excited radical, Ar2CH'**, by electron transfer, there is a reduction with increasing [n-butyl chloride]. Unless a change of mechanism from electron transfer to transfer of a chlorine atom is invoked, thedecreasecan beunderstoodin termsof increasingcontributions of in-cage ion collapse (decreasing escape of ions) as the polarity of the solvent decreases as AN is replaced by n-butyl chloride. This interpretation is in agreement with the increased yield of permanent ground-state radical depletion.

Summary and Conclusions The key speciesin the photoionizationmechanism is the excited state radical, Ar2CH'* (DI state)?.6bv46 which is produced from the ground-state radical, Ar2CH', by absorption of one photon of, e.g., 308- or 248-nmI8light. Ar2CH** (lifetime in AN 100330 ns) is chemically inert with respect to methyl chloride,n-butyl chloride, and dichloromethane, but it reacts with CC14, but not to give cation. Ar2CH**is able to absorb a photon of, e.g., 308or 248-nmI8light. This process leads to ionization. The ionization yields (as measured at 1 2 0 ns) can be increased by n-BuC1, but not by CH2Cl2 or CC14. The result with n-BuCI indicates that in the initial cation-(solvated) electron pair there is some electron return (to yield radical) which can be intercepted by dissociative electron capture giving C1-. At higher concentrations of n-BuC1, the negative effect of the decreasing solvent polarity overcompensates that of the increased electron scavenging capacity of the solvent. Scheme I represents a summary of the energetics of excitation and ionization of Ph2CH'. The ionization potential of Ph2CH* is 7.32 eV.I3 If it is assumed that the solvation energy of the electron in AN is 1.5 eV, about half4' of that (3.17 eV)I8of C1-, and if the solvation energy of Ph2CH+ is taken as 2.91 eV,l8 it is evident that ionization of Ph2CH*with only one 308 nm photon (quantum energy 4.0 eV) is thermodynamically possible. In contrast, without any solvation, ionization is still uphill, even after absorption of a second 308 nm photon (by DI. The assumption is made that after absorption of the 308 nm quantum by PhzCH', relaxation to DI (see Scheme I) is very fast, k > 10I2 s-l). For the system to profit from solvation, that process has to occur on the same timescale as ionization. Solvation times of organic molecules in AN have been determined to be 0.5 to 1 much longer than those of electron movement (